孤立性肺结节MSCT征象的多因素回归分析对良恶性肺结节鉴别诊断的价值研究

发布时间：2018-04-08 11:09

本文选题：孤立性肺结节　切入点：x线计算机　出处：《辽宁医学院》2013年硕士论文

【摘要】：目的通过对孤立性肺结节的各种影像征象的研究，并运用回归分析的方法，对孤立性肺结节的良恶性进行鉴别诊断，并得出相应的结论，以提高CT检查方法对良恶性肺结节的鉴别诊断，从而给临床治疗提供更加可靠的依据。方法用回顾性方法分析188例经病理确证获得证实的孤立性肺结节的临床及CT资料。收集各病例临床资料并分析肺结节影像学征象，结合临床和影像的数据进行统计学分析来判定不同影像学特征对良恶性结节的诊断手段的可靠性。影像学特征包括结节的大小，所在部位，分叶征，，毛刺征，血管集束征，胸膜凹陷征，钙化和空泡征。良恶性结果为因变量，影像学征象为自变量。建立数据进行多因素回归分析。从这些自变量中计算出恶性结节的危险因素，建立估算模型。然后通过模型方程得到单个SPN的概率值。并将结果与临床实际诊断结果进行对比。进一步描绘出ROC曲线，并最终判断逻辑回归分析对SPN良恶性诊断的应用价值。结果经过统计188例孤立性肺结节中有恶性结节共106例，包括腺癌63例，小细胞癌20例，鳞癌16例，大细胞癌5例，细支气管癌3例；良性结节共82例，包括38例结核球，35例炎性病变，7例错构瘤，2例血管瘤。其中左肺上叶7例，左肺下叶56例，右肺上叶6例，右肺中叶26例，右肺下叶93例。本文中研究的CT征象统计比例为；恶性结节分叶征69.8%、毛刺征60.4%、胸膜凹陷征67.0%、血管集束征74.5%、空泡征25.5%，钙化8.5%。在良性肺结节组中出现率依次为分叶征20.7%,毛刺征23.2%,胸膜凹陷征17.1%,血管集束征13.4%,空泡征7.3%,钙化39.0%。良性结节多边缘光滑，长毛刺的出现率较高，结节内钙化的发生率等与恶性结节组相比较，差异均具有统计学意义(P0.05)。另外，患者临床症状、发病年龄及发病部位对做出孤立性肺结节的诊断和鉴别诊断亦有重要的参考依据。结论孤立性肺结节的CT征象的分析对良恶性的鉴别诊断具有一定的参考价值。其中毛刺征，分叶征，血管集束征，胸膜凹陷征，空泡征提示恶性可能性较大。钙化提示良性结节可能性较大。运用统计学回归分析计算出OR值和95%可信区间并绘出其ROC曲线，可以明确该价值的临床意义，为临床最终做出确切诊断提供可靠的依据。
[Abstract]:PurposeThrough the study of various imaging signs of solitary pulmonary nodules and the method of regression analysis, the differential diagnosis of benign and malignant solitary pulmonary nodules was carried out, and the corresponding conclusions were drawn.In order to improve the CT diagnosis of benign and malignant pulmonary nodules, and provide a more reliable basis for clinical treatment.MethodThe clinical and CT data of 188 cases of solitary pulmonary nodules confirmed by pathology were analyzed retrospectively.The clinical data of each case were collected and the imaging signs of pulmonary nodules were analyzed. The reliability of different imaging features in the diagnosis of benign and malignant nodules was evaluated by statistical analysis combined with clinical and imaging data.Imaging features include nodule size, location, lobulation, burr, vascular cluster, pleural indentation, calcification and vacuolation.The benign and malignant results were dependent variables and the imaging signs were independent variables.The data were analyzed by multivariate regression analysis.The risk factors of malignant nodules were calculated from these independent variables and an estimation model was established.Then the probabilistic value of a single SPN is obtained by the model equation.The results were compared with the results of clinical diagnosis.The ROC curve was further described, and the value of the logistic regression analysis in the diagnosis of benign and malignant SPN was finally determined.ResultThere were 106 cases of malignant nodules in 188 solitary pulmonary nodules, including 63 cases of adenocarcinoma, 20 cases of small cell carcinoma, 16 cases of squamous cell carcinoma, 5 cases of large cell carcinoma, 3 cases of bronchiolar carcinoma, 82 cases of benign nodules.There were 38 cases of tuberculous bulbus tuberculum 35 cases of inflammatory lesions 7 cases of hamartoma and 2 cases of hemangioma.There were 7 cases of left upper lobe, 56 cases of left lower lobe, 6 cases of right upper lobe, 26 cases of right middle lobe and 93 cases of right lower lobe.The proportion of CT signs studied in this paper is: malignant nodule lobulation sign 69.8, burr sign 60.4, pleural sag sign 67.0, vascular cluster sign 74.5, vacuole 25.5and calcification 8.5.In the benign pulmonary nodule group, the occurrence rates were as follows: lobulation sign 20.7m, burr sign 23.2m, pleural indentation sign 17.1m, vascular cluster sign 13.4m, vacuole sign 7.3g and calcification 39.0.The incidence of calcification in benign nodules was significantly higher than that in malignant nodules (P 0.05).In addition, the clinical symptoms, onset age and location of the patients also have important reference for the diagnosis and differential diagnosis of solitary pulmonary nodules.ConclusionThe analysis of CT findings of solitary pulmonary nodules has certain reference value for differential diagnosis of benign and malignant lung nodules.Among them, burr sign, lobulation sign, vascular cluster sign, pleural indentation sign and vacuole sign suggest that malignancy is more likely.Calcification suggests that benign nodules are more likely.Using statistical regression analysis to calculate OR value and 95% confidence interval and draw its ROC curve can determine the clinical significance of this value and provide a reliable basis for the final diagnosis.
【学位授予单位】：辽宁医学院
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：R563

【参考文献】